improve documentation about CONDBAT (& MDBAT) and how to watch the limit being approached

The IBM provided documentation for important ZPARM of CONDBAT and MDBAT is a bit vague on exactly what to watch for these important limits. The documentation could be improved.

I reviewed the documentation for DISPLAY DDF command

https://www.ibm.com/docs/en/db2-for-zos/12?topic=commands-display-ddf-db2

It display the important ZPARM of CONDBAT and MDBAT (zparm name of MAXDBAT)

These two values are “limits” and rarely change. The DISPLAY DDF shows us these ZPARM values to remind us of these limits.

The DISPLAY DDF DETAIL command gives us all the information to watch how we are approaching these limits but it is not obviously stated in the documentation (in my opinion).

MDBAT limit is relatively better documented then CONDBAT which really beats about the bush on what exactly to watch.

(1)

The DISPLAY DDF DETAIL command page (link above) should have clear statement that

“DISPLAY DDF DETAIL message DSNL090I shows MDBAT (value of system parameter MAXDBAT). Assuming DT=A, then watch DSNL092I ADBAT value as it approaches MDBAT. When ADBAT reaches MDBAT then connections will begin to queue while waiting to become active”

“DISPLAY DDF DETAIL message DSNL090I shows CONDBAT (value of system parameter CONDBAT). Assuming DT=A, then watch DSNL092I & DSNL093I. When the total of “INACONN +ADBAT-DSCDBAT” reaches CONDBAT then incoming connections cannot be made and the distributed application will experience a connection error (connection refused)”

(2)

In addition, or alternatively, you could make enhance the documentation for the actuall messages to be more clear for how to watch CONDBATT

https://www.ibm.com/docs/en/db2-for-zos/12?topic=messages-dsnl092i

The message documentation for DSNL092i does say for ADBAT that if MAXDBAT is reached then queuing occurs. In my opinion, the documentation does not emphasize enough that this is important… but OK… you say something

The message documentation for DSNLO93I does not say how to watch CONDBAT. I find the documentation vague on this important point! All the information is here… but it is not clearly stated

Why is this important to me?

We had an experience when CONDBAT was reached. It turns out that our CONDBAT was relatively low and the application introduced a new change that forgot to properly close connections… so we hit CONDBAT. We later raised CONDBAT and changed the application to properly close connections. Both actions addressed our problem at that point in time. My point is that we hit CONDBAT and it was not immediately obvious to me during this application incident. It was not immediately obvious to me what I needed to be checking and monitoring. Of course, I did figure it out… but it took a bit to read and re-read and really understand the information from DISPLAY DDF DETAIL

As further background. I opened IBM support CASE TS005127784 to ask IBM Support directly and clearly how to monitor CONDBAT. You can see that it took IBM support some thought to provide the proper answer.